Brain metastasis occurs in 25% of cancer patients, and 50% of brain tumors are of metastatic origin. Molecular mechanisms responsible for the development of brain metastasis are poorly understood, mainly due to the lack of relevant animal models. We have developed one of the first approaches to accurately detect dormant tumor cells and the development of brain metastasis from a single cell. We describe an efficient xenograft model of brain metastasis initiated from hematogenously delivered single cells. The model is based upon DU145 human prostate cancer cells that were originally isolated from a brain metastasis. Although the parental cells are non-metastatic in mouse models, activation of the RasV12G37 effector signaling pathway leads to the formation of highly vascularized brain metastasis. We couple this model with a supra-paramagnetic iron oxide micron particle (MPIO)-enhanced magnetic resonance imaging (MRI) detection of single cells. The goal of this project is to allow us to investigate the three possible fates of cancer cells following their arrival in the brain: cell death, the establishment of dormant solitary cells/small pre-angiogenic metastasis, or the development of large vascularized metastasis. Initially, it has been necessary to establish the validity of this model. MPIOs were shown to have no demonstrable toxicity to DU145 tumor cells in vitro or in vivo. To establish the clonal origin of brain metastases, graded mixtures of GFP-labeled and RFP-labeled DU145 cells were administered to mice, and the distribution of fluorescently labeled brain metastases was quantified. Only singly labeled tumors were observed, and the distribution of GFP or RFP-labeled tumors reflected the original proportion of such cells in the inoculums. MPIOs appear as hypo intense regions on MRI scans. Several avenues of investigation have been used to verify that hypo intense regions observed on MR images correlated with MPIO-labeled tumors. Longitudinal collection of MR images following systemic introduction of MPIO-labeled tumor cells showed that 50-100 tumor cells/100,000 cells introduced into the arterial blood arrest in the brain, and subsequently, about 5% of these cells develop into brain metastasis. Line scan analyses demonstrated significant hypo intensity that could not be attributed to background intensity variations or anatomical variation (i.e. blood vessels) but correlated with regions of metastatic tumor development. Dissemination maps relative to time following systemic introduction of tumor cells were generated with doubly fluorescent and MPIO-labeled tumor cells using coupled MRI and histological analyses. Vascular and intra-parenchyma distributions were analyzed following FITC-dextran administration in order to label blood vessels. This system is particularly valuable for investigating biological mechanisms that contribute to the initiation of metastases from single cells and the effect of targeted therapies on such processes. We have evaluated the effects of blocking VEGF-A activity. We have observed that blocking VEGF-A activity two weeks after the establishment of tumor cells in the brain resulted in significantly fewer and smaller metastasis at 5 weeks as determined by MRI-derived volumetric analyses. Ongoing studies are directed at determining the effects of blocking VEGF-A activity on metastatic efficiency and tumor dormancy of DU145(RasV12G37) cells following their arrest in the brain. Histological and MRI-determined tumor blood volume analyses also are being performed.